Hi all,

as I told earlier I am doing mp3 streaming using  high level api of MAD decoder for buffering I am using the "streamnextframe" which is in the frame.c file and  for the beginning frame address I am using "framebegin "  which is in the stream.c . By using these two variables I am filling the buffer in the input callback .with this  concept I am able to decode the song .with some errors messages are comming as bad data pointer ,forbidden bitrate value etc.........and while playing in the winamp some little noise and some of the song words are some what changed( Pronunciation is different but perceptually identifiable )at every 20 sec or 30 sec ( this time duration is  varying some times) .whether my concept of  filling the buffer with those variables are correct or not .next problem is I am converting the output file into WAV  file using the software and at the same time I am giving the same input file to the other soft wares which will convert directly from MP3  to WAV . then what I feel was   the file size I am getting is less compared to the other one (software decoded). why is it so .where is the problem . I am some what suspicious above concept. whether that is correct or not .if it is not what i have to do and how I have to make links with the decode engine in such a way that i can able to play the song with buffering concept. 

thanks in advance

regards,

prafulla chandra